A lot more oversight is needed to ensure omics-based technologies are effective and ready for clinical use, an Institute of Medicine (IOM) report concluded last week. Particularly when results from omics-based tests are used to devise therapies and medical interventions for patients, researchers and healthcare institutions should be sure that the tests have been proven scientifically valid.
The National Cancer Institute commissioned the report after omics-based tests developed by Duke University researchers were found to be based on faulty and fraudulent statistical models. The problematic tests were used in three clinical trials for determining the most effective chemotherapy treatment for lung and breast cancer patients. The scandal has since led to multiple paper retractions as well as lawsuits from patients involved in the trials.
A strengthening of oversight at multiple levels—from the investigators and their institutions, to the US Food and Drug Administration and scientific journals—will prevent similar cases from happening in the future, according to the IOM. “Many of these failures stemmed from problems that may exist at other institutions,” the report’s authors wrote.
For example, omics-based tests intended to inform decisions about patients’ medical care are legally required to undergo review by the FDA before they are used in clinical trials—a requirement that Duke University mistakenly overlooked, Nature reported.
The report also outlines ways in which researchers can overcome the problem of “overfitting”—or the incorrect linking of an observed omics signature with a clinical outcome.